Many children with specific language impairment (SLI) show serious limitations in grammatical ability. An especially common manifestation of this grammatical weakness is the inconsistent use of grammatical morphemes that mark tense and agreement (e.g., -s in runs and is in The girl is crying). This inconsistency serves as a reliable means of distinguishing children with SLI from their typically developing peers. Although the details of this inconsistent use have been well described in the scientific literature, the factors responsible for this hallmark symptom are far from clear. This lack of understanding has prevented the development of successful intervention methods for ameliorating this problem. In this application, we propose that tense and agreement inconsistency stems from these children's failure to understand the structural dependencies that hold within a wide variety of sentences that appear in the input, and, as a result, this inconsistency does not require an assumption of a separate learning mechanism that has gone awry. The common feature of the problematic sentences is that a nonfinite subject-verb sequence appears late in the sentence, and is allowed because finiteness is expressed in a verb form that appears earlier in the sentence. Examples include We saw the girl/her running, Is the girl/she crying?, and Did the girl/she finish her breakfast? As a result of the children's failure to grasp these structural dependencies, they often extract nonfinite subject-verb sequences (e.g., The girl/her running, The girl/she crying) and use these inappropriate forms as the basis for generating new utterances with the same structure. Although emerging evidence suggests that these children do, in fact, extract nonfinite subject-verb sequences from larger structures, we have needed a procedure that allows us to test whether these children are sensitive to the relationship between early-appearing finite forms in the sentence and later-appearing nonfinite subject-verb sequences. Through application of the looking-while-listening paradigm, we can now begin to test this proposal. In four experiments, we will examine the shifts in eye gaze of preschool-age children with SLI and their typically developing peers as they hear questions while looking at pairs of pictures on a video screen. We expect that typically developing children will orient toward the target picture on the basis of early-appearing finite information (e.g., shifting towarda picture of a plural referent upon hearing the auxiliary are in Are the nice little cats eating?). I contrast, we expect that children with SLI will fail to make use of this information and instead orient toward the target picture only after the plural noun (e.g., cats) is heard. Furthermore, we expect that children's sensitivity or insensitivity to this early-appearing finite information will serve as a significant predictor of their consistency/inconsistency in using tense and agreement finite forms in their own speech. Findings that support this proposal would have significant implications for prevailing theory and suggest a wholly different approach to treating this persistent grammatical problem in children with SLI.